Chapter
Fairness and Demographic Properties in AI
25:20 - 36:24 (11:04)
The episode discusses fairness guarantees for demographic properties such as race, gender, and age in AI systems, especially in the context of criminal sentencing and parole decisions, which remain unresolved issues in the debate over affirmative action.
Clips
Discussions surrounding fairness guarantees at the individual level simultaneously tackle group fairness for all possible combinations of groups.
25:20 - 31:28 (06:08)
Summary
Discussions surrounding fairness guarantees at the individual level simultaneously tackle group fairness for all possible combinations of groups. Fairness discussions do not get politicized along the lines of race and gender when demographic properties are not biological, but can still be strong determinants of outcomes and personality.
ChapterFairness and Demographic Properties in AI
EpisodeMichael Kearns: Algorithmic Fairness, Bias, Privacy, and Ethics in Machine Learning
PodcastLex Fridman Podcast
Discussions on fairness within the criminal justice system don't come without their debates, as they tend to focus solely on the criminal element involved, begging the question - what about victims?
31:28 - 36:24 (04:56)
Summary
Discussions on fairness within the criminal justice system don't come without their debates, as they tend to focus solely on the criminal element involved, begging the question - what about victims? The goal is often to reach zero unfairness and error, as well as the same false incarceration rates for people of all races.